Seventy years have elapsed since Maxwell Finland published articles describing bacterial resistance to antimicrobial agents at the Boston City Hospital. Reference Finland1 Since then, the evolution and dissemination of resistance has emerged as a global “crisis” and “calamity.” Reference Neu2,Reference Kunin3 However, the distinction between the genetics conferring resistance and their incidence rates may be blurred. Simply put, the recognition of a new mechanism of resistance does not tell us about the burden of that organism. Longitudinal surveillance data can provide long-term trends in the incidence of these multidrug-resistant organisms (MDROs) and help to assess whether local control strategies are effective and sustainable. For example, despite the fear of a post-antibiotic era, when incidence rates for 6 MDROs in a large cohort in the United States were calculated over a 6-year span, trends were decreasing for 4 of them. Reference Jernigan, Hatfield and Wolford4 The accompanying editorial found “recent evidence of progress” and “some success.” Reference Fang and Schooley5
Twenty-four years ago, we designed a prospective surveillance system to describe the incidence rates of 6 MDROs among hospitalized adults in our medical center. We have continued to collect data with the same methodology and the same personnel. This method yields prospective, monthly data which can be used to assess the efficacy of our infection prevention, diagnostic stewardship, and antimicrobial stewardship programs. This work extends our findings and presents 24 years of data. We hypothesize that such data are necessary to fully describe the epidemiology of MDROs and assess the success or failure of control strategies.
Methods
Setting
The University of Vermont Medical Center is a 499-bed, tertiary-care, academic medical center and a teaching hospital for the University of Vermont Larner College of Medicine. There is a hospital epidemiologist, 6 infection prevention practitioners, a director of antimicrobial stewardship, and an infectious disease pharmacist.
Surveillance method
The method has been described previously. Reference Ahern and Alston6–Reference Ahern, Noyes, Smith and Alston8 It was developed in 2001 and generates monthly incidence rates for 6 MDROs. It is laboratory-based and includes both colonization and infection for (1) methicillin-resistant Staphylococcus aures (MRSA), (2) vancomycin-resistant Enterococcus (VRE), (3) Clostridioides difficile, (4) fluoroquinolone-resistant Pseudomonas aeruginosa, (5) ceftazidime-resistant gram-negative bacilli, and (5) Stenotrophomonas maltophilia. Positive cultures or nucleic acid amplification tests collected from hospitalized adults >48 hours after admission were included. Pediatrics, rehabilitation, psychiatry, and patients with cystic fibrosis were excluded. The numerator is the number of patients with one of the 6 MDROs, and the denominator is hospital-wide patient-days. Patients with a given MDRO are only counted once.
Visual inspection of the hospital-wide incidence of all 6 organisms combined suggested that the 24 years of data could be divided into 3 time periods: 2001–2010, 2011–2017, and 2018–2024. See Figure 1. Incidence rate ratios were compared with Poisson regression. Stata version 19.5 software was used (StataCorp, College Station, TX).
Figure 1. Incidence rate for 6 multidrug-resistant organisms over 24 years at an academic medical center. The data are divided into 3 time periods: 2001–2010, 2011–2017, and 2018–2024. A. Clostridioides difficile. B. Methicillin-resistant Staphylococcus aureus. C. Vancomycin-resistant Enterococcus. D. Quinolone-resistant Pseudomonas aeruginosa. E. Ceftazidime-resistant gram-negative bacilli. F. Stenotrophomonas.
Results
The 24 years were divided into 3 time periods: 2001–2010, 2011–2017, and 2018–2024. The number of patient-days (denominator) for the 3 time periods ranged from 572,611 to 832,221. The number of patients with one of the 6 MDROs (numerator) ranged from 906 to 2,884. The incidence rates for all 6 organisms combined during the 3 time periods were 3.47, 1.98, and 1.28/1,000 patient-days. Using the first 10 years as a reference, the rate ratios for the second and third periods were 1.75 and 2.71 times lower, respectively (P < .01).
The figure demonstrates how the numbers of the 6 individual MDROs contributed to the overall rate. For example, the incidence rate for Clostridioides difficile fell from 1.04 to 0.59/1,000 patient-days (P < .01), and MRSA fell from 0.86 to 0.18/1,000 patient-days (P < .01). Remarkably, the rate of quinolone-resistant Pseudomonas aeruginosa fell from 0.46 to 0.03/1,000 patient-days (P < .01). In fact, the incidence rates for all 6 MDROs fell during the 2nd and 3rd time periods when compared to the first (all 12 comparisons with P < .01 except ceftazidime-resistant gram-negative bacilli (P = .05).
Discussion
This work describes a surveillance system we have created and maintained for nearly a quarter century. It demonstrates that for selected MDROs, sustained reductions in incidence rates are possible.
Strengths of this work include 24 years of consistent data collection using the same definitions and performed by the same person. Outlier months caused by short-lived and unsustainable interventions are smoothed out by the longitudinal design. Which MDROs to choose can be dictated by local risk assessments. Data collection is not overly time consuming or costly. Patients counted in the numerator can be assigned to wards to allow unit-specific feedback. The incidence rates can be correlated with antimicrobial usage down to the ward level.
Limitations of this work are that the data come from a single center. The 6 MDROs included were chosen in 2001 and may not all be the ones we would choose if we were starting today. This design is not useful for detecting newly emerged pathogens since you must select which organisms to include at the outset. Finally, the trends revealed by longitudinal data require a commitment of time.
Space does not allow us to list the actions that were implemented during the 3 time periods. Our data suggest that clinically meaningful and statistically significant reductions in the incidence rates of MDROs can be achieved with sustained, evidence-based, commonsense approaches. These data provide important evidence that transmission of MDROs can be slowed in an academic medical center. Transmission is central to microbial evolution. This study provides a glimmer of hope.
Acknowledgements
The authors would like to thank the Infection Prevention Team at the University of Vermont Medical Center.
Financial support
No financial support was provided relevant to this article.
Competing interests
All authors report no conflicts of interest relevant to this article.
